BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:MechE Colloquium: Landscape and generalisation in deep learning
DTSTART:20200922T121500
DTEND:20200922T131500
DTSTAMP:20260408T034138Z
UID:db7ab4c76b5a80c9fc9a7e880cb9ef74aeeba0ba407dd6cce5e70262
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. Matthieu Wyart\, Physics of Complex Systems Laboratory\,
  EPFL School of Basic Sciences (SB)\, Institute of Physics (IPHYS)\nIf you
  would like to attend the talk in BM 5202\, please register here (on a fir
 st-come\, first-served basis). This allows us to limit the number of peopl
 e in the room and to satisfy contact tracing requirements.\n\nFor remote a
 ttendance: Zoom link\n\n\nAbstract:\nDeep learning is very powerful at a v
 ariety of tasks\, including self-driving cars and playing go beyond human 
 level. Despite these engineering successes\, why deep learning works remai
 ns unclear\; a question with many facets.  I will discuss two of them: (i
 ) Deep learning is a fitting procedure\, achieved by defining a loss funct
 ion which is high when data are poorly fitted.  Learning corresponds to a
  descent in the loss landscape. Why isn’t it stuck in bad local minima\,
  as occurs when cooling glassy systems in physics? What is the geometry of
  the loss landscape? (ii) in recent years it has been realised that deep l
 earning works best in the over-parametrised regime\, where the number of f
 itting parameters is much larger than the number of data to be fitted\,  
 contrarily to intuition and to usual views in statistics.  I will propose
  a resolution of these two problems\, based on both an analogy with the en
 ergy landscape of repulsive particles and an analysis of asymptotically wi
 de nets.\n\nBio:\nMatthieu Wyart is Associate Professor in the Institute o
 f Physics at EPFL in Switzerland. He received his B.A. at École Polytechn
 ique and obtained his Ph.D. degree from CEA\, Saclay. He was a George Carr
 ier Fellow at Harvard University before joining the Physics Department at 
 NYU in 2010. He became Associate Professor in 2014 and moved to EPFL in 20
 15.\nOne focus of Wyart's work is the classification of the elementary exc
 itations controlling the linear and the plastic response in amorphous mate
 rials. He introduced the notion that some of these excitations are margina
 lly stable in the solid phase. Such marginality fixes key aspects of the s
 tructure\, and implies that the density of excitations presents a pseudo-g
 ap. These notions are important to describe the low-temperature properties
  of glasses\, the elasticity near the jamming transition\, the rheology of
  dense granular and suspension flows\, the yielding transition in foams or
  metallic glass\, and more generally glassy systems with sufficiently long
 -range interactions. Wyart's work has also focused on the neuronal circuit
  of simple organisms. He received a Sloan Fellowship in 2011 and became a 
 Simons Investigator in 2015.
LOCATION:BM 5202 https://plan.epfl.ch/?room==BM%205202
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR